From: Hal Fulton <hal9000@...>
Date: 2004-05-03T09:41:44+09:00
Subject: Re: Strange behaviour of Strings in Range

Yukihiro Matsumoto wrote:
> Hi,
> 
> In message "Re: Strange behaviour of Strings in Range"
>     on 04/05/03, Hal Fulton <hal9000@hypermetrics.com> writes:
> 
> |Will Rite have a better-defined notion of "character"? Perhaps including
> |Unicode and such?
> 
> No.  The definition of "character" should belong to the application
> domain, I believe.  Considering internationalization, any particular
> definition of character can not satisfy all.

These are two or three separate issues, I believe.

I know that no one encoding scheme will suffice for Asian languages
as well as European. Unicode in that sense is largely a dream, as I
understand it.

And I do not favor a Char class, which seems unnecessary to me.

But here are some related questions, to get more specific:

1. Will str[0] always be a Fixnum?

2. Will ?x always be a Fixnum?

3. In addition to each_byte, would each_char make sense? As I see it,
it would default to be the same as each_byte, but would be replaced
for a wide-char or multibyte variable-length encoding.


But I18N is one of the areas of my greatest ignorance in Ruby.


Thanks,
Hal